In this paper, we analyze how a weighted voting system introduced in 1862, which shifted the distribution of political power from landowners to industrialists at local town meetings, affected investments in local public education. We use an event study design based on a newly constructed panel data set with annual observations of nearly 2,200 Swedish local governments over 28 years, i.e., more than 60,000 observations. Most importantly, there is no pre-trend in educational spending in our event study but rather a sharp change in the dynamic treatment effects exactly at the date when the treatment occurs, i.e., when industrialists receive more political power at town meetings. The estimated cumulative treatment effect is also economically substantial. For example, per capita spending on education increased by approximately 37% within 6 years in local governments where industrialists came to political power. Our findings are therefore consistent with the idea that political institutions are a key determinant of human capital accumulation and long-run economic development. • This work is financed via an ERC consolidator grant 616496 (Pettersson-Lidbom).